This study utilized data from the first and second waves of the Malaysian Elders Longitudinal (MELoR) study. The measurement of pruritus, psychological status, sleep quality and self-rated health are described below.
Sample population
The MELoR study is a cohort study comprising of community-dwelling older adults aged 55 years or over at the time of recruitment between 2012 and 2015. Participants, in the first wave study, were selected through simple random sampling stratified by age and ethnicity from the electoral rolls of the parliamentary constituencies of Petaling Jaya North, Petaling Jaya South, and Pantai Valley. This study recruited 1623 participants from the community. Two hundred seventy-four with incomplete data were excluded and the remaining 1349 participants were included for follow-up. However, 45 died and 534 were unable to follow up due to insufficient resources, therefore only 770 completed the follow-up assessments in the second wave study. The study methods have been described in detail elsewhere [12]. This study was approved by the University of Malaya Medical Centre Medical Ethics Committee (Ref: 925.4) and complied with the Helsinki Declaration of 1975, revised in 1983. Written informed consent was obtained from all study participants prior to their inclusion.
Data collection
First wave data were collected through a detailed home-based computer-aided questionnaire interview followed by a hospital visit for clinical assessments. Information on basic demographics, socioeconomic status, living environment, health, psychology, quality of life, function, social participation, lifestyle and health beliefs were obtained at baseline. Eligible participants were then interviewed again in the second wave between 2015 and 2016. Changes in health, psychological, functional, social, lifestyle, and quality of life outcomes were measured using a computerised questionnaire survey in the second wave, as in the first wave. Second wave data were obtained entirely from a home-based questionnaire survey administered by the trained researchers.
Changes in health, psychological, functional, social, lifestyle, and quality of life outcomes were measured using a computerised questionnaire survey. Psychological symptoms were assessed with the 21-item Depression, Anxiety and Stress Scale (DASS-21), perceived health status using self-rated health, and sleep quality suing the Pittsburgh Sleep Quality Index (PSQI). For the purpose of this study, medical history from both the first wave and second wave were merged to establish prevalence of each condition at the second wave. Only DASS-21 results from the second wave was included in this analysis.
The DASS-21 consists of seven items each for each domain of depression, anxiety and stress rated on a four-point Likert scale, with 0 indicating “did not apply to me at all”, and 3 indicating “applied to me all the time”. The sum of scores are calculated for each domain, and the total score is then multiplied by two. The maximum score for each domain is therefore 42, with 0 indicated absence of any symptoms and 42 indicating extremely severe symptom. Symptoms of depression, anxiety, or stress symptoms were considered present for those with score of 10 or above, 8 and above and 15 and above for the respective domains [13].
The Pittsburgh Sleep Quality Index was used to assess the quality and patterns of sleep by evaluating seven components of sleep. It contains 19 self-rated questions and 5 questions rated by the bed partner or roommate (if available). Only self-rated questions are included in the scoring. The 19 self-rated items are combined to form seven “component” scores, each of which has a range of 0–3 points. In all cases, a score of “0” indicates no difficulty, while a score of “3” indicates severe difficulty. The component scores were summed to produce a global score ranging from 0 to 21. A score of 5 or higher indicated poor sleep quality [14].
Self-rated health (SRH) was assessed using the two questions, “Would you say your health is……?” and “In general, compared to other people your age, would you say your health is…….? The possible responses were: “poor”, “fair”, “good”, “very good” and “excellent”. Response for each question was then dichotomized to “poor or fair” and “good, very good or excellent”.
Pruritus
Participants were asked at the second wave whether they had experienced pruritus in the preceding week. Those who responded ‘yes’ to the presence of pruritus in the preceding week were then asked whether the pruritus was localised or generalised. They were also asked to rate the severity of pruritus on a scale on 0–10 using a numerical rating scale (NRS) which represents a row of 10 rectangles with increasingly heaving shading starting from no shading for 0–1 and black solid shading from 9 to 10. Numerical numbers from 0 to 10 were also included beneath the margin between each rectangular box, and at each end. The NRS is a unidimensional scale which has similarly been validated for measuring the severity of pain, where 0 represents “having no symptoms” while ten represents “having the worst imaginable symptoms.” For NRS, a score of 0 to 2 represents no or mild pruritus, a score of 3–6 represents moderate pruritus, while a score of 7 to 8 and 9 to 10 represents severe and very severe pruritus, respectively [15]. For chronic pruritus, the NRS, visual analogue scale (VAS) and verbal rating scale (VRS) have shown high reliability and concurrent validity (r > 0.8; p < 0.01) [16].
Statistical analysis
Data were analysed using the SPSS version 21.0 (IBM Corporation). Essential characteristics obtained at wave one was then compared against those with and without pruritus. For parametric data, the independent t-test was used to compare means while the Mann-Whitney U test was used to compare medians for non-parametric data. The Pearson chi-squared test or Fisher’s exact test where necessary were used to determine possible factors associated with generalised pruritus. All tests were two-tailed with a significance level of p < 0.05. Potential confounders were adjusted for by using multivariate logistic regression with stepwise backward method. The fit of the logistic models was assessed on the basis of the Hosmer-Lemeshow test and p < 0.05 (two tailed) was considered as statistically significant.